Intrinsic Activity of Metal Centers in Metal-Nitrogen-Carbon Single-Atom Catalysts for Hydrogen Peroxide Synthesis

Article abstract of DOI:10.1021/jacs.0c10636

Metal-nitrogen-carbon (M-N-C) single-atom catalysts (SACs) show high catalytic activity for many important chemical reactions. However, an understanding of their intrinsic catalytic activity remains ambiguous because of the lack of well-defined atomic str

Full text of DOI:10.1021/jacs.0c10636

pubs.acs.org/JACS
Article
Intrinsic Activity of Metal Centers in Metal−Nitrogen−Carbon
Single-Atom Catalysts for Hydrogen Peroxide Synthesis
Chang Liu,∥ Hao Li,∥ Fei Liu,∥ Junsheng Chen, Zixun Yu, Ziwen Yuan, Chaojun Wang, Huiling Zheng,
Graeme Henkelman,* Li Wei,* and Yuan Chen*
Cite This: https://dx.doi.org/10.1021/jacs.0c10636
Read Online
sı
*
Metrics & More
Article Recommendations
Supporting Information
ACCESS
ABSTRACT: Metal−nitrogen−carbon (M−N−C) single-atom catalysts
(SACs) show high catalytic activity for many important chemical reactions.
However, an understanding of their intrinsic catalytic activity remains
ambiguous because of the lack of well-defined atomic structure control in
current M−N−C SACs. Here, we use covalent organic framework SACs with
an identical metal coordination environment as model catalysts to elucidate
the intrinsic catalytic activity of various metal centers in M−N−C SACs. A
pH-universal activity trend is discovered among six 3d transition metals for
hydrogen peroxide (H2O2) synthesis, with Co having the highest catalytic
activity. Using density functional calculations to access a total of 18 metal
species, we demonstrate that the difference in the binding energy of O2* and
HOOH* intermediates (EO * − EHOOH*) on single metal centers is a reliable
2
thermodynamic descriptor to predict the catalytic activity of the metal centers.
The predicted high activity of Ir centers from the descriptor is further validated experimentally. This work suggests a class of
structurally defined model catalysts and clear mechanistic principles for metal centers of M−N−C SACs in H2O2 synthesis, which
may be further extendable to other reactions.
with distinct catalytic properties. Further, carbon matrixes with
different doped atoms, chemical functional groups, or defects
may also have high catalytic activities.22−25 Alternatively,
organometallic complexes containing M−N−C active sites
have been studied as homogeneous ORR catalysts for
decades.26−29 Tetradentate N coordination sites can firmly
anchor metal atoms.30 Previous studies have assembled
porphyrin- or phthalocyanine-based organometallic complexes
on graphene or carbon nanotubes as M−N−C SACs.31−33
Although metal coordination sites in these catalysts are
identical, graphene or carbon nanotube substrates induce
charge transfer to M−N−C active sites, which is expected to
significantly alter their catalytic activities.34 Carbon materials in
these catalysts can also contribute to the observed catalytic
activities. On the other hand, various theoretical descriptors
have been proposed to explain the catalytic activity of M−N−
C SACs, such as intermediates adsorption energies,13,35−38
metal d-band centers,39 and metal−crystal field stabilization
energies.40 However, because existing M−N−C SACs lack
1. INTRODUCTION
Single-atom catalysts (SACs) typically refer to heterogeneous
catalysts with isolated single metal atoms embedded within a
solid matrix. These catalysts have attracted considerable
interest in recent years due to their intriguing properties,
including high metal atom utilization efficiency, controlled
coordination environments of metal atoms, unique quantum
size effects, and tunable metal−support interactions.1−6 In
particular, metal−nitrogen−carbon (M−N−C) SACs (where
M is a metal atom coordinated with four N or C atoms within
a carbon skeleton) have shown excellent catalytic activities for
oxygen reduction reaction (ORR).7−9 For example, Fe−N−C
SACs are highly active for the 4e− ORR (O2 + 4H+ + 4e− →
2H2O, E0 = 1.23 V) in converting O2 to H2O,10,11 which is an
essential reaction for fuel cells and metal−air batteries.12,13
Co−N−C SACs show a high selectivity toward the 2e− ORR
(O2 + 2H+ + 2e− → H2O2, E0 = 0.70 V) to produce H2O2,14,15
which is an imperative chemical for many applications.16,17
M−N−C SACs have been synthesized by either bottom-up
methods, such as atomic layer deposition, wet chemistry
synthesis via binding metal ions to carbon matrixes, or ball-
milling, as well as top-down methods, including pyrolysis of
metal-containing complexes, metal−organic frameworks, poly-
mers, and small organic precursors, or some solid-state
reactions.18−21 Metal atoms in M−N−C SACs resulting from
these synthesis methods are often at unique coordination sites
Received: October 6, 2020
https://dx.doi.org/10.1021/jacs.0c10636
J. Am. Chem. Soc. XXXX, XXX, XXX−XXX
© XXXX American Chemical Society
A

Journal of the American Chemical Society
pubs.acs.org/JACS
Article
Materials Characterization. Scanning electron microscopy
(SEM) images were collected on a Zeiss Ultra Plus microscope.
High-angle annular dark-field scanning transmission electron
microscopy (HAADF-STEM) images were collected on an FEI
Themis-Z microscope. Fourier transform infrared (FTIR) spectra
were collected on a Thermo Scientific Nicolet 6700 spectrometer
using the attenuated total reflection mode. X-ray diffraction (XRD)
patterns were obtained on a PANalytical X’Pert diffractometer using a
Cu Kα source (λ = 1.5406 Å). N2 physisorption isotherms were
collected on a Quantachrome Autosorb iQ2 gas sorption analyzer.
Atomic force microscopy (AFM) height profiles were obtained on a
Park NX10 microscope using the noncontact mode. The samples
were acid-digested in HNO3 for 6 h before inductively coupled
plasma atomic emission spectroscopy (ICP-AES) elemental analysis
on a Varian Vista Pro instrument. X-ray photoelectron spectroscopy
(XPS) spectra were collected on a Thermo Fisher K-Alpha+
spectrometer equipped with an Al Kα source (1486.3 eV). X-ray
absorption spectroscopy (XAS) spectra were collected in fluorescence
mode on the BL-12B2 beamline at the National Synchrotron
Radiation Research Center, Taiwan. The energy was tuned by a
double-crystal Si (111) monochromator. Data analysis and fitting
were performed with the Demeter package using the FEFF 9.0 code.
Electrochemical Measurements. Catalysts were dispersed in a
1/9 water/isopropanol (v/v) solution (with 0.05 wt %/v Nafion 117)
at a concentration of 1 mg mL−1 by bath sonication for 30 min. The
catalyst ink was then drop-casted on prepolished glassy carbon disk of
rotating ring-disc electrodes (RRDEs) with a Pt ring electrode (Pine
Instrument, disk o.d. of 5.5 mm, ring i.d. of 6.5 mm, and o.d. of 8.5
mm) with a mass loading of 5 μg cm−2. Electrochemical measure-
ments were conducted on a CHI 760E electrochemical workstation in
a three-electrode configuration with a graphite rod (Pine) as the
counter electrode and an Ag/AgCl (sat. KCl) electrode as the
reference electrode at 25 °C. All potentials were subjected to a
reversible hydrogen electrode (RHE) by adding 0.197 + pH × 0.059
V. Three types of O2-saturated electrolytes were used, including 0.1 M
KOH (pH = 12.6), 0.1 M phosphate-buffered saline (PBS) (pH =
7.4), and 0.1 M sodium acetate solution (ABS) (pH = 3.5). Before
commencing ORR performance measurements, RRDEs were
precycled with 20 cyclic voltammetry (CV) scans between 0.1 and
1 V versus RHE. Linear sweep voltammetry (LSV) polarization
measurements were performed under a scan rate of 2 mV s−1 with a
rotation speed of 1600 rpm without iR correction. The potential of
the Pt ring was kept at 1.2 V versus RHE. The onset potentials are
defined as the potential required to reach a geometric current density
of −0.05 mA cm−2 for the disk electrode (Udisk) and 0.02 mA cm−2 for
the ring electrode (Uring), respectively. The influence of SCN−
adsorption was assessed by collecting LSV polarization curves in
O2-saturated 0.1 M ABS with or without adding 0.1 M KSCN. ORR
electron transfer number (n) is determined by the following equation,
well-defined and uniformly distributed metal centers without
added interferences from various other catalytic active species,
their precise structure−activity relationships in M−N−C SACs
remain unclear.37,41,42
Here, we unambiguously reveal the intrinsic electrocatalytic
activity of various metal centers in M−N−C SACs for H2O2
synthesis using two-dimensional (2D) covalent organic
framework (COF)-based model catalysts (denoted as COF-
366-M). The conjugated porphyrin-based COF (COF-366) is
a coordination 2D polymer crystal with a periodically repeating
architecture.43 Its abundant porphyrin moieties can host a
variety of 3d transition metals or noble metals with an identical
chemical structure. Further, their conjugated polymer frame-
work provides sufficient electron-transfer capability without the
need to add carbon material substrates, while the 2D porous
nanosheet structure enables efficient mass transfer. Our
experimental studies show that Co centers have the highest
activities among 3d transition metals, including Mn, Fe, Co,
Ni, Cu, and Zn, in three types of electrolytes at different pH
values. Using density functional theory (DFT) calculations, we
show that the binding energy difference between O2* and
HOOH* intermediates (EO * − EHOOH*) at single metal sites
2
can be used as a general descriptor to predict the intrinsic
catalytic activity of these metal centers, including ten 3d
transition metals and eight noble metals, in COF-366-M.
Experimental results further corroborate the prediction of
COF-366-Ir as a highly active catalyst.
2. METHODS
Chemicals and Materials. 5,10,15,20-(Tetra-4-aminophenyl)
porphyrin (TAPP, 98%) was purchased from PorphyChem, Inc.
Terephthaldehyde (TPD, 99%), Mn(OAc)2 (99.99%), Fe(OAc)2
(99.99%), Co(OAc)2·4H2O (99.99%), Ni(OAc)2·4H2O (99.99%),
Cu(OAc)2 (99.99%), Zn(OAc)2·2H2O (99.99%), IrCl3·H2O (99.9%),
Na2SO4 (anhydrous, >99%), KOH (99.9%, semiconductor grade),
NaOAc (99%), KH2PO4 (99%), K2HPO4 (99%), CeSO4·4H2O
(98%), methylene blue (97%), acetic acid (99.9%), KSCN (99%),
methanol (anhydrous, 99.9%), ethanol (200 proof, anhydrous,
>99.5%), N,N-dimethylformamide (DMF, anhydrous, 99.8%),
mesitylene (98%), chloroform (>99.5%), benzonitrile (anhydrous,
99%), and H2O2 solution (30 wt %) were purchased from Sigma-
Aldrich. Ar (5.0 grade) and O2 (4.5 grade) gases were purchased from
BOC Australia.
Synthesis of TAPP-M. In a typical synthesis, 0.3 mmol TAPP
(200 mg) and 1.2 mmol metal acetate salt were suspended in 20 mL
of methanol before the addition of 90 mL of chloroform and 30 mL of
DMF. The mixture was stirred under Ar protection at 80 °C for 24 h.
After cooling to room temperature, the solution was transferred into a
separatory funnel and washed thoroughly with deionized (DI) water
(18.2 MΩ, Millipore, 3 × 100 mL). The organic layer was collected
and dried over anhydrous Na2SO4, and the product was further
vacuum-dried overnight to obtain the final product. For the synthesis
of TAPP-Ir, a mixture of 0.035 mmol TAPP (23 mg) and 0.14 mmol
IrCl3·H2O (50 mg) was mixed in 50 mL of benzonitrile. The mixture
was refluxed for 24 h under Ar protection. Afterward, the procedure is
the same as that used for other TAPP-M.
idisk
n = 4 ×
idisk + iring/N
(1)
where idisk and iring are the currents obtained from the glassy carbon
disk and Pt ring, respectively. N is the calibrated collection efficiency
of RRDEs (N = 0.38). The selectivity toward H2O2 formation can be
evaluated by two methods: Faradaic efficiency (λFaradaic, %) or the
fraction of O2 reduced to H2O2 (λH O , %).
2
2
iring
N
λFaradaic
=
× 100
Synthesis of COF-366 and COF-366-M. They were synthesized
via the imine condensation in solvothermal reactions. TAPP or
TAPP-M (0.02 mmol) and 0.04 mmol TPD were suspended in a
mixture of 1 mL of absolute ethanol, 1 mL of mesitylene, and 0.1 mL
of 6 M acetic acid in a Pyrex tube by sonication. After sonication for
15 min, the tube was flash-frozen at 77 K (liquid N2 bath), evacuated
to an internal pressure of 150 mTorr, and then flame-sealed. After
reaction at 120 °C for 72 h, the dark purple solids were collected by
vacuum filtration and washed with dimethylacetamide (DMAC) and
absolute ethanol before drying under vacuum.
idisk
(2)
2 × iring/N
idisk + iring/N
λH O
=
× 100
2
2
(3)
The following equations were used to calculate the product-specific
current density for H2O2 formation (jH O ) and H2O formation (jH O),
2
2
2
j
= iring/(N × Ageom)
H2O2
(4)
B
https://dx.doi.org/10.1021/jacs.0c10636
J. Am. Chem. Soc. XXXX, XXX, XXX−XXX

Journal of the American Chemical Society
pubs.acs.org/JACS
Article
Figure 1. Synthesis and structural characterization of COF-366-M. (a) Schematic illustration of the synthesis of COF-366-M. (b, c) AFM images of
COF-366 and COF-366-Co and the corresponding height profiles along the red lines; the scale bar is 500 nm. (d) XRD patterns. (e) N2
physisorption isotherms. (f) Pore-size distribution. (g) High-resolution N1s XPS spectra. (h) High-resolution Co 2p XPS spectra. (i) Fitted Co K-
edge extended X-ray absorption fine structure (EXAFS) spectrum of COF-366-Co. The inset shows a HAADF-STEM image. The scale bar is 5 nm.
j
= j − j
disk H2O2
where j is the measured current density and jL and jK are the diffusion-
limiting and kinetic current densities, respectively. ω is the angular
velocity, n is electron transferred number, F is the Faraday constant,
C0 is the saturated concentration of O2 in electrolytes at room
temperature, D0 is the diffusion coefficient of oxygen in the
electrolytes, and υ is the kinematic viscosity of the electrolyte at 25
°C. Tafel analysis was performed by using the extracted jK values. The
mass-specific current density (jK‑mass, mA mg−1 cm−2) and the
turnover frequency (TOF) of the catalysts were calculated by using
the following equations,
H2O
(5)
where Ageom is the surface area of the Pt ring.
The kinetic limiting current density jK was extracted from the
polarization curves using the K−L equation,
1
j
1
1
1
1
jK
=
+
=
+
Bω1/2
jL
jK
(6)
(7)
B = 0.62nFC0(D0)2/3v−1/6
C
https://dx.doi.org/10.1021/jacs.0c10636
J. Am. Chem. Soc. XXXX, XXX, XXX−XXX

Journal of the American Chemical Society
pubs.acs.org/JACS
Article
jK × λH O × Adisk
corresponding to (100), (200), (111), and (310) facets of
COF-366, respectively, consistent with a simulated eclipsed
stacking model.52 No other metal or metal oxide XRD peaks
are observed. The XRD results indicate that the incorporation
of 3d-metals does not affect the crystal structure of COF-366.
N2 physisorption isotherms (shifted in the y-axis) of COF-366-
M and COF-366 in Figure 1e are all type I isotherms. The
Brunauer−Emmett−Teller (BET) specific surface areas of
COF-366-M are comparable from 998 to 1117 m2 g−1 (Table
S1). Figure 1f shows that they also have similar pore-size
distributions calculated by the DFT method, agreeing with the
porous structure shown in their physical model in Figure
1a.52,53
Both XPS (Figure S5) and ICP-AES confirm that the
abundance of 3d metals in COF-366-M is similar at around
0.77 to 0.84 at% (Table S1). C 1s XPS spectra (Figure S6)
suggest that COF-366-M have identical chemical structures in
their C atoms. N 1s XPS spectra (Figure 1g and Figure S7)
display that the position (at 398.4 eV) and abundance of N in
imine bonds are the same in COF-366 and COF-366-M.
However, pyrrolic-N in porphyrin rings (at 399.4 eV in COF-
366) shifts to a higher binding energy in COF-366-Co due to
the interactions between N and Co. The deconvolution of
metal 2p XPS spectra (Figure 1h and Figure S8) indicates that
most of the 3d transition metals incorporated into COF-366-M
are in their divalent state, except that Co in COF-366-Co and
Fe in COF-366-Fe are in mixed divalent and trivalent states.
We also collected the electron paramagnetic resonance (EPR)
of Co in COF-366-Co (Figure S9). The featureless EPR
spectrum suggests that Co is at a low-spin state (S = 0),
promoting Co-active sites’ electrochemical activity.54 Because
the redox potential of Co2+/3+ and Fe2+/3+ is 1.25 V and 0.75 V
versus RHE, respectively, both Co and Fe would be reduced to
their divalent states under the applied potential used for
ORR.37,55
We further analyzed the representative COF-366-Co by
XAS. Its k3-weighted Fourier-transformed extended X-ray
absorption fine structure (FT-EXAFS) spectrum at the Co
K-edge was fit over the first and second Co shells. Figure 1i
shows a sharp peak at 1.54 Å originating from Co−N bonds,
while the peaks from Co−C bonds in the porphyrin rings in
the second shell are weak, indicating that Co atoms are isolated
in the porphyrin. The catalysts were further examined by
atomic-resolution STEM conducted in HAADF mode. The
inset of Figure 1i and other images in Figure S10 show isolated
bright spots of single metal atoms without metal aggregates.
Overall, the physicochemical characterization results confirm
that all COF-366-M catalysts have similar physical structures
and metal abundance, and the incorporated 3d transition
metals are present in the same chemical environment in the
porphyrin.
Experimental Evaluation of the Catalytic Activity of
COF-366-M. We evaluated the electrocatalytic performance of
COF-366-M for ORR by the RRDE method (Figure S11 and
the related description). Three types of O2-saturated electro-
lytes at different pH values were used, including alkaline 0.1 M
KOH (pH = 12.6), neutral 0.1 M PBS (pH = 7.2), and acidic
0.1 M ABS (pH = 3.5) electrolytes. Catalysts were drop-cast
on a prepolished and calibrated glassy carbon disk electrode
that has a collection efficiency of 0.38 (Figure S12).56 The
catalyst mass loading on the disk electrodes was also
optimized; 5 μg cm−2 was used in this study (Figure S13
and the related discussion). Figure 2a and Figure S14 show
2
2
jK‐mass
=
(8)
(9)
0.005
jK × λH O × Adisk
2
2
TOF =
n × m × w × F/Mw
where 0.005 (mg cm−2) is the catalyst loading. Adisk is the geometric
surface area of the disk electrode, n = 2 is the electron-transfer
number for H2O2 formation, m is the mass of the catalyst loaded on
the disk electrode, w is the metal composition by weight, and Mw is
the molar weight of the metal. F is the Faraday constant. H2O2RR
tests were carried out by subtracting LSV currents measured in Ar-
saturated 0.1 M ABS containing 1 mM H2O2 from the background
LSV currents measured in Ar-saturated 0.1 M ABS without H2O2.
DFT Calculations. All DFT calculations were performed with the
Vienna ab initio simulation package. Valence electrons were described
by the Kohn−Sham wave functions expanded in a plane-wave basis
set.44 Electron correlation was described by a generalized gradient
approximation method and the Perdew−Burke−Ernzerhof (PBE)
functional.45,46 Core electrons were described by a projector
augmented-wave method.47 Geometries were defined as converged
when all the force on each atom fell below 0.05 eV Å−1. The Brillouin
zone was sampled at the Γ-point. Spin-polarization was considered in
all the calculations. The kinetic barriers were calculated by the
climbing image nudged elastic band (Cl-NEB) method, with at least
six images located between the initial and final states.48 Entropic
corrections (at a temperature of 298 K) and zero-point energies were
applied in the free energy calculations, with the values taken from a
previous study.49 The binding energies of O2 (EO *) and HOOH
(EHOOH*) were calculated using
2
E
= Etot − Ebare − E
O2*
O2
(10)
(11)
EHOOH* = Etot − Ebare − EHOOH
where Etot is the total energy of the system with adsorbate, Ebare is the
total energy of the bare system, EO is the total energy of an O2
2
molecule in a vacuum, and EHOOH is the total energy of an H2O2
molecule in a vacuum. To test the sensitivity of the applied functional,
similar calculations using the revised Perdew−Burke−Ernzerhof
(RPBE) functional were also performed.50 Except for a slight
systematic shift in the O2 binding energies compared to PBE
(∼0.18 0.13 eV), no significant change was found in the adsorption
configuration.
3. RESULTS AND DISCUSSION
Synthesis and Structural Characterization of COF-
366-M. We synthesized COF-366-M with an imine con-
densation reaction.51 As illustrated in Figure 1a, 5,10,15,20-
(tetra-4-aminophenyl) porphyrin salts loaded with different
3d-transition metal ions (TAPP-M, M = Mn, Fe, Co, Ni, Cu,
and Zn) react with TPD in solvothermal reactions (see details
in the Methods section). COF-366 without metals was also
synthesized as a reference. New imine bond vibrations at vC=N
of 1621 and 1192 cm−1 in FTIR confirm the successful
formation of COF-366 (Figure S1). As-synthesized COF-366-
M and COF-366 self-assemble into microspheres with a
diameter of ∼500 nm (see SEM images in Figures S2 and S3).
These microspheres can be easily delaminated into 2D
nanosheets by simple bath-sonication in organic solvents
(e.g., N,N-dimethylformamide or N-methyl-2-pyrrolidone).
Parts b and c of Figure 1 show AFM images of delaminated
COF-366 and COF-366-Co nanosheets with thicknesses of ∼2
nm. The other COF-366-M catalysts exhibit similar
morphologies in their delaminated nanosheets (Figure S4).
XRD patterns of COF-366-M catalysts (Figure 1d) exhibit
identical peaks at 2θ of 3.47°, 6.96°, 8.66°, and 12.60°,
D
https://dx.doi.org/10.1021/jacs.0c10636
J. Am. Chem. Soc. XXXX, XXX, XXX−XXX

Journal of the American Chemical Society
pubs.acs.org/JACS
Article
Figure 2. Electrocatalytic performance of COF-366-M. (a) ORR RRDE polarization curves of COF-366 and COF-366-M in O2-saturated 0.1 M
KOH (pH = 12.6). (b) H2O2 selectivity in three types of electrolytes. (c) TOF of COF-366-M catalysts. (d) Catalytic activity of COF-366-M
catalysts for the H2O2RR.
polarization LSV curves of COF-366-M collected from the disk
and ring electrodes in the three types of O2-saturated
electrolytes. All tested COF-366-M catalysts exhibit higher
disk current densities (jdisk) and more positive disk onset
potentials (Udisk) than COF-366, indicating that the metal
atoms in COF-366-M are active catalytic sites. As tabulated in
Table S2, COF-366-Fe has the highest activity for ORR,
delivering the largest jdisk of 2.90 mA cm−2 at 0.1 V versus RHE
and the most positive Udisk at 0.803 V versus RHE. The overall
trend for the ORR activity is COF-366-Fe > Ni > Mn ≈ Co >
Cu > Zn.
different COF-366-M catalysts, as displayed in Figure 2b,
exhibit a volcano-shaped curve. As summarized in Table S3,
COF-366-Co has the highest λH O of 91% and λFaradaic of 84%
2
2
in 0.1 M KOH electrolyte, exhibiting a high selectivity toward
H2O2 formation by the 2e− transferred ORR. In comparison,
COF-366-Mn is more selective toward 4e− ORR to produce
H2O. We further extracted jK of different catalysts and
calculated the mass-specific H2O2 current density (jK‑mass
,
Figure S16). COF-366-Co can deliver a jK‑mass of nearly 977
mA cm−2 mg−1 at 0.2 V versus RHE. Figure 2c shows the H2O2
production turnover frequency (TOF) of the metal centers in
COF-366-M. COF-366-Co delivers a superior TOF of 1.79 s−1
at 0.4 V versus RHE, which is ∼1.3 times higher than COF-
366-Ni at 1.37 s−1. COF-366-Co maintains a high H2O2
selectivity in a wide potential window, and its TOF reaches
9.05 s−1 at 0.1 V versus RHE. The Tafel slopes of COF-366-M
catalysts in the three electrolytes were calculated from their jK.
Figure S17 and Table S4 revealed similar Tafel slopes,
suggesting that they have the same rate-limiting step in the
different electrolytes.57
The iring in the upper panel in Figure 2a and Figure S14 were
compared to evaluate the selectivity toward H2O2 formation.
Results tabulated in Table S2 show that COF-366-Co delivers
the highest iring of 0.199 mA at 0.1 V versus RHE in 0.1 M
KOH, followed by COF-366-Ni at 0.168 mA. The overall
selectivity trend toward H2O2 formation is COF-366-Co > Ni
> Cu > Fe > Zn > Mn in 0.1 M KOH. Similar patterns were
observed in both neutral and acidic electrolytes. We further
compared the selectivity toward H2O2 of COF-366-M catalysts
quantitatively, based on λH O and the electron-transfer number
2
2
We further evaluated the H2O2 production capability of the
(n, Figure S15). The values of λH O at 0.3 V versus RHE of
optimal COF-366-Co catalyst in an H-shaped electrolyzer
2
2
E
https://dx.doi.org/10.1021/jacs.0c10636
J. Am. Chem. Soc. XXXX, XXX, XXX−XXX

Journal of the American Chemical Society
pubs.acs.org/JACS
Article
Figure 3. DFT calculation results of ORR on COF-366-M. (a) Optimized geometry of COF-366-Co. (b) Charge density difference before and
after incorporating Co in the porphyrin moiety. (c−h) Geometries of adsorbed reaction intermediates and adsorption energies on COF-366-Co at
U = 0 V. (c, f) ORR via the 2e− pathway toward H2O2 formation, (d, g) ORR via the OOH dissociation 4e− pathway, and (e, h) ORR via the
HOOH dissociation 4e− pathway.
(Figure S18). The electrolyzer can operate steadily in ambient
air, delivering a gradually increasing H2O2 productivity (Figure
S19a) as quantified by a Ce(SO4)2 titration method (Figure
S20). The COF-366-Co catalyst can produce H2O2 up to 909
and their polarization curves were collected in an Ar-saturated
0.1 M ABS electrolyte containing 1 mM H2O2. The
background-corrected jdisk in Figure 2d indicates that the
activity for H2O2RR agrees with the selectivity toward H2O2
obtained in the ORR tests discussed earlier. COF-366-Co is
almost inert to H2O2RR with a negligible measured jdisk. In
contrast, COF-366-Fe and COF-366-Mn are highly active for
H2O2RR. They also tend to adsorb H2O2 molecules and
further break the O−O bonds to produce H2O.
mmol gcat h−1 at a current density of 22 mA cm−2 with a
−1
device Faradaic efficiency (λdevice) of 79%. The catalyst stability
was evaluated by a 3 h chronopotentiometric test performed at
22 mA cm−2 (Figure S19b). The operating potential remained
stable, while the cumulative H2O2 concentration (cH O
)
2
2
Theoretical Insights on the Catalytic Activity of COF-
366-M. DFT calculations were used to understand the
catalytic activity of metal centers in COF-366-M. Figure 3a
shows the optimized geometry of COF-366-Co. Figure 3b
displays the calculated charge density difference before and
after incorporating a Co atom into the porphyrin ring, which
indicates a significant charge redistribution with an electron
delocalized from the metal center to the conjugated 2D COF.
The direct O2 dissociation is less favorable on the metal site
because two O* (* stands for the active site) cannot be stably
coadsorbed. The free energy of the reaction intermediates on
COF-366-M was calculated by the computational hydrogen
electrode method.49 Figure 3c illustrates the ORR along the
2e− transferred pathway on COF-366-M. ORR may also
proceed via the 4e− transferred pathways by either OOH
dissociation (Figure 3d) or H2O2 dissociation (Figure 3e). The
increased linearly. Negligible morphological changes were
observed under HADDF-STEM and Co 2p XPS spectrum
after the stability test (Figure S21). The cH O reached 337 mM
2
2
after the 3-h test, which is sufficient to be utilized directly for
wastewater remediation via Fenton reaction. As a demon-
stration, we added 5 mL of the catholyte into 10 mL of
acidified methylene blue solution (200 mg L−1, pH = 1, with
0.1 mmol Fe2+, Figure S22a). After a brief hand-shaking, the
color of the solution quickly faded (Figure S22b). UV−vis
absorption measurements further confirmed the complete
removal of the dye (Figure S22c).
H2O2 produced from ORR may be further reduced to water
via the H2O2 reduction reaction (H2O2RR) by the same
catalyst, which would sabotage H2O2 production. We
compared the catalytic activity of COF-366-M for H2O2RR,
F
https://dx.doi.org/10.1021/jacs.0c10636
J. Am. Chem. Soc. XXXX, XXX, XXX−XXX

Journal of the American Chemical Society
pubs.acs.org/JACS
Article
Figure 4. Validation of the thermodynamic descriptor (EO * − EHOOH*) on COF-366-M catalysts. (a) Correlation between EO * − EHOOH* and
2
2
ORR performance parameter, O2 efficiency (λH O ), in the three electrolytes. Correlation between EO * − EHOOH* and (b) onset potential for ring
2
2
2
electrode (Uring), (c) specific current density for H2O2 formation (jH O ) at 0.3 V versus RHE, and (d) specific current density for H2O formation
2
2
(j ) at 0.3 V versus RHE. (e) Calculated E and EHOOH* on different metal centers in COF-366-M. (f) Comparison between E − EHOOH*
H2O
O2*
O2*
and experimentally measured λH O for various COF-366-M catalysts.
2
2
G
https://dx.doi.org/10.1021/jacs.0c10636
J. Am. Chem. Soc. XXXX, XXX, XXX−XXX

Journal of the American Chemical Society
pubs.acs.org/JACS
Article
corresponding free energies of reaction intermediates via the
three ORR pathways are listed in Table S5 and also illustrated
in parts f, g, and h of Figure 3, respectively. HOOH* is
unstable on COF-366-Mn, prohibiting the formation of H2O2,
which agrees with its lowest H2O2 selectivity observed in our
experiments. Hence, COF-366-Mn was excluded from the
calculations along the 2e− transferred pathway (Figure 3f) and
the H2O2 dissociation pathway (Figure 3h). Further, the
adsorption of two OH* on COF-366-Zn is also not stable, and
thus it was also excluded from the H2O2 dissociation pathway
(Figure 3h). We further analyzed the kinetics of OOH*
dissociation by the Cl-NEB method (Figure S23). We found
that all of these systems have relatively large OOH*
dissociation barriers (Table S6); in good agreement with our
experimental results, 4e− ORR is less facile at these COF-336-
M. Meanwhile, because HOOH* dissociation on most of the
COF-336-M is endothermic (Figure 3h), it is expected that
this step also has relatively large kinetic barriers, which in turn
leads to difficulty in the following 4e− ORR step. Although the
scaling rule between the free energies of OOH* and OH* has
been widely used to explain the reaction mechanism of ORR
on closely packed metal surfaces,17,49,58 we did not observe the
same scaling relation on COF-366-M. Figure 3 suggests that
the 2e− pathway toward H2O2 formation is favored over the
4e− ORR pathways on COF-366-M, which agrees with the
experimental observation that λH O is >50% on all COF-366-
related to the formation and subsequent desorption of the
HOOH* reaction intermediate. As we can see from the
calculated EHOOH* on different systems, most of them have a
moderate binding to the metal center with the absolute
HOOH binding energies slightly below zero. Therefore, an
ideal center metal at COF-366 is expected to have moderate
binding energies for both O2* and HOOH*.
Parts a−d of Figure 4 show correlations between the EO *
−
2
EHOOH* descriptor and four experimental performance
parameters in the three different electrolytes, including the
H2O2 efficiency (λH O ), the onset potential for ring electrode
2
2
(Uring), the specific current density for H2O2 formation (jH O ),
2
2
and the specific current density for H2O formation (jH O), both
2
obtained at 0.3 V versus RHE. Each of these correlations shows
a volcano-shaped curve for the different COF-366-M catalysts.
COF-366-Co and COF-366-Ni sit on the top of the curves of
the catalytic performance parameters, while their EO *
−
2
EHOOH* approaches zero. Therefore, we can set a clear target
(EO * − EHOOH* ≈ 0 eV) to rapidly screen promising COF-
2
366-M catalysts according to this empirical descriptor. Note
that this new descriptor is the electronic energy; the use of free
energy would lead to the same conclusion since the entropic
terms among the different SAC systems are similar. We also
plotted λH O against several commonly used thermodynamic
2
2
2
2
descriptors (Figure S26)60 and EO * or EHOOH* alone (Figure
M catalysts (Table S3).
2
Our DFT calculations and experimental results indicate that
the selectivity of ORR toward H2O2 formation on COF-366-M
is predominately determined by the binding strength of O2 and
HOOH intermediates on the metal sites. First, Figure 3
suggests that the adsorption of an oxygen molecule to form
O2* on metal atom sites could be the key step of H2O2
formation. Our calculation results (Table S5) show that O2
adsorption on Cu or Zn is relatively weak, while it is strong
enough on Co, Ni, and Fe. The effect of O2 adsorption was
experimentally validated by comparing LSV polarization curves
in the ABS electrolyte with and without 0.1 M SCN−. SCN−
can strongly bind to single metal atom sites, which would block
that metal site for O2 adsorption.59 Figure S24 shows that the
onset potential for ORR (Udisk) increases significantly on all
COF-366-M with the addition of SCN− in the electrolyte. We
can establish an explicit correlation between the change in the
S27); none of the other descriptors show a better correlation
with our experimental data.
The close agreement between our experimental and
theoretical results suggests that the volcano-like trend for
selective H2O2 formation among the tested COF-366-M
catalysts may be used with the EO * − EHOOH* descriptor to
2
predict the catalytic performance of other transition metals in a
similar molecular structure, i.e., M−N4/C in porphyrin-like
molecules. We computed EHOOH* and EO * on single-atom sites
2
in 12 other COF-366-M catalysts, in which M includes four
other early 3d transition metals (Sc, Ti, V, and Cr) and eight
different noble metals (Ru, Rh, Pd, Ag, Os, Ir, Pt, and Au). The
results are displayed in Figure 4e and Table S8. EO * on the
2
early 3d transition metals is much stronger (i.e., more
negative) than EHOOH*, suggesting that the O−O bond cannot
be easily retained at these sites, resulting in unfavorable
conditions for H2O2 formation. The revised Perdew−Burke−
Ernzerhof (RPBE) functional was also used to test the
sensitivity of the applied functional. Table S8 shows that,
except for a slight systematic shift in the O2 binding energies
compared to PBE (∼0.18 0.13 eV), no significant change
was found in the adsorption configuration.
onset potential (ΔUdisk) and the binding energy of O2* (EO *
)
2
on COF-366-M catalysts (Figure S25 and Table S7). Second,
although O2 adsorption on COF-366-Fe has the lowest
adsorption energy, COF-366-Fe is favorable for HOOH*
decomposition rather than HOOH* desorption (Figure 3h).
Experimental results in Figure 2d also show that COF-366-Fe
has one of the highest jdisk values for H2O2RR, which suggests
that the binding strength of HOOH on the metal site is also
critical for H2O2 formation. Overall, the optimal COF-366-M
catalyst should bind to these two intermediates, neither too
strongly nor too weakly.
We overlapped the EO * − EHOOH* descriptor on the
2
experimental results in Figure 4f to show that COF-366-M
incorporated with early 3d metal centers are not predicted to
be active for H2O2 formation. There is no clear trend among
COF-366-M incorporated with noble metals. Ir stands out
among the noble metals considered with a near-zero binding
energy difference between O2* and HOOH*, suggesting that
COF-366-Ir may be selective for H2O2 synthesis. On the basis
of this theoretical prediction, we synthesized a COF-366-Ir
catalyst (details are provided in the Methods section). AFM
measurements (Figure S28) suggested that the COF-366-Ir
exhibits an identical nanosheet morphology to other COF-366-
A New Thermodynamic Descriptor to Predict
Catalytic Activity. On the basis of these findings, we propose
that the difference between the binding energies of O2* and
HOOH* (EO * − EHOOH*) may be used as a thermodynamic
2
descriptor to describe the catalytic activity of COF-366-M
catalysts for H2O2 formation. An intermediate EO * is needed
2
to promote the initial adsorption of oxygen molecules and
meanwhile inhibit the activation of the O−O bond. EHOOH* is
H
https://dx.doi.org/10.1021/jacs.0c10636
J. Am. Chem. Soc. XXXX, XXX, XXX−XXX

Journal of the American Chemical Society
pubs.acs.org/JACS
Article
M samples. The formation of single-atom Ir sites was
confirmed by HAADF-STEM, XPS, and EXAFS analyses
(Figure S29). The Ir abundance in COF-366-Ir was
determined by ICP-AES at 0.81 at%, similar to other COF-
366-M. The catalytic performance of COF-366-Ir was
evaluated, and the results are shown in Figure S30. COF-
366-Ir exhibits λH O of 83% and λFaradaic of 71% in a 0.1 M
KOH electrolyte, which fits well with the experimentally
measured catalytic activity trend observed for other COF-366-
M catalysts incorporated with 3d transition metals (Figure 4f).
The University of Texas at Austin, Austin, Texas 78712,
United States; orcid.org/0000-0002-0336-7153;
Email: henkelman@utexas.edu
Li Wei − School of Chemical and Biomolecular Engineering,
The University of Sydney, Darlington, New South Wales
2006, Australia; orcid.org/0000-0001-8771-2921;
Email: l.wei@sydney.edu.au
Yuan Chen − School of Chemical and Biomolecular
Engineering, The University of Sydney, Darlington, New
South Wales 2006, Australia; orcid.org/0000-0001-
9059-3839; Email: yuan.chen@sydney.edu.au
2
2
4. CONCLUSIONS
Authors
In summary, we synthesized a series of 2D COF-based M−N−
C SACs, in which different metals can be incorporated at
porphyrin moieties with the same structure and catalytic active
site density. These COF-366-M catalysts allowed us to
compare and understand the catalytic activity and selectivity
of different metal centers for H2O2 synthesis via ORR in
alkaline, neutral, and acidic electrolytes. COF-366-Co is the
most active with a high λH O at 91%, λFaradaic at 84%, a large
Chang Liu − School of Chemical and Biomolecular
Engineering, The University of Sydney, Darlington, New
South Wales 2006, Australia
Hao Li − Department of Chemistry and the Oden Institute for
Computational and Engineering Sciences, The University of
Texas at Austin, Austin, Texas 78712, United States;
orcid.org/0000-0002-7577-1366
Fei Liu − State Key Laboratory of Applied Microbiology
Southern China, Guangdong Provincial Key Laboratory of
Microbial Culture Collection and Application, Guangdong
Institute of Microbiology, Guangdong 510070, P. R. China
Junsheng Chen − School of Chemical and Biomolecular
Engineering, The University of Sydney, Darlington, New
South Wales 2006, Australia
Zixun Yu − School of Chemical and Biomolecular Engineering,
The University of Sydney, Darlington, New South Wales
2006, Australia
Ziwen Yuan − School of Chemical and Biomolecular
Engineering, The University of Sydney, Darlington, New
South Wales 2006, Australia; orcid.org/0000-0002-
3331-0668
Chaojun Wang − School of Chemical and Biomolecular
Engineering, The University of Sydney, Darlington, New
South Wales 2006, Australia
Huiling Zheng − Department of Chemistry and the Oden
Institute for Computational and Engineering Sciences, The
University of Texas at Austin, Austin, Texas 78712, United
States; orcid.org/0000-0001-8347-3724
2
2
TOF of 9.05 s−1 per Co site, and productivity of 909 mmol
gcat h−1, which is one of best among recently reported
−1
catalysts (Table S9). Combing experimental and computa-
tional approaches, we show that the initial O2 adsorption and
the stability of HOOH* intermediate govern the activity and
selectivity toward H2O2 formation collaboratively. The binding
energy difference between the O2* and HOOH* intermediate
(EHOOH* − EO *) can be used as a thermodynamic descriptor
2
to evaluate the catalytic performance of different metal centers
in COF-366-M, including both 3d transition metals and noble
metals. Other than Co and Ni, Ir was predicted to have a high
catalytic performance, which was subsequently verified
experimentally. These findings shed new light on the rational
design of M−N−C SACs of precisely controlled structures for
ORR and beyond. The thermodynamic descriptor established
here may be extendable to other ORR catalysts with well-
defined structures.
ASSOCIATED CONTENT
* Supporting Information
The Supporting Information is available free of charge at
https://pubs.acs.org/doi/10.1021/jacs.0c10636.
■
sı
Complete contact information is available at:
https://pubs.acs.org/10.1021/jacs.0c10636
Additional experimental results including FTIR spectra,
SEM images of COF-366 and COF-366-M, additional
AFM images of COF-366-M, XPS survey scans, C 1s, N
1s, and M2p spectra, HAADF-STEM images of COF-
366-M, RRDE calibration curves, catalyst mass loading
optimization, ORR performance, TOF values and Tafel
curves of COF-366-M in electrolytes of different pH
values, ORR electrolyzer diagram and H2O2 production
performance, dye removal, poststability test STEM and
XPS results, SCN− poisoning test, correlations between
experimental performance and theoretical descriptors,
and characterization results and ORR performance of
COF-366-Ir; and tabulated structural characterization
results and detailed computation data (PDF)
Author Contributions
∥C.L., H.L., and F.L. contributed equally.
Notes
The authors declare no competing financial interest.
ACKNOWLEDGMENTS
■
This work is supported by the Australian Research Council
under the Future Fellowships scheme (FT160100107).
Calculations at UT were supported by the Welch Foundation
(F-1841) and the Texas Advanced Computing Center.
REFERENCES
■
(1) Yang, X. F.; Wang, A. Q.; Qiao, B. T.; Li, J.; Liu, J. Y.; Zhang, T.
Single-atom catalysts: a new frontier in heterogeneous catalysis. Acc.
Chem. Res. 2013, 46 (8), 1740−1748.
AUTHOR INFORMATION
Corresponding Authors
Graeme Henkelman − Department of Chemistry and the
Oden Institute for Computational and Engineering Sciences,
■
(2) Zhu, C. Z.; Fu, S. F.; Shi, Q. R.; Du, D.; Lin, Y. H. Single-atom
electrocatalysts. Angew. Chem., Int. Ed. 2017, 56 (45), 13944−13960.
(3) Wang, A. Q.; Li, J.; Zhang, T. Heterogeneous single-atom
catalysis. Nat. Rev. Chem. 2018, 2 (6), 65−81.
I
https://dx.doi.org/10.1021/jacs.0c10636
J. Am. Chem. Soc. XXXX, XXX, XXX−XXX

Journal of the American Chemical Society
pubs.acs.org/JACS
Article
(4) Liu, L. C.; Corma, A. Metal catalysts for heterogeneous catalysis:
from single atoms to nanoclusters and nanoparticles. Chem. Rev. 2018,
118 (10), 4981−5079.
(5) Jiao, L.; Jiang, H. L. Metal-organic-framework-based single-atom
catalysts for energy applications. Chem. 2019, 5 (4), 786−804.
(6) Gawande, M. B.; Fornasiero, P.; Zboril, R. Carbon-based single-
atom catalysts for advanced applications. ACS Catal. 2020, 10 (3),
2231−2259.
(7) Wang, Y.; Mao, J.; Meng, X.; Yu, L.; Deng, D.; Bao, X. Catalysis
with two-dimensional materials confining single atoms: concept,
design, and applications. Chem. Rev. 2019, 119 (3), 1806−1854.
(8) Lu, B. Z.; Liu, Q. M.; Chen, S. W. Electrocatalysis of single-atom
sites: impacts of atomic coordination. ACS Catal. 2020, 10 (14),
7584−7618.
(9) Shi, Z. S.; Yang, W. Q.; Gu, Y. T.; Liao, T.; Sun, Z. Q. Metal-
nitrogen-doped carbon materials as highly efficient catalysts: progress
and rational design. Adv. Sci. 2020, 7 (15), 2001069.
(24) Lu, Z.; Chen, G.; Siahrostami, S.; Chen, Z.; Liu, K.; Xie, J.;
Liao, L.; Wu, T.; Lin, D.; Liu, Y.; Jaramillo, T. F.; Nørskov, J. K.; Cui,
Y. High-efficiency oxygen reduction to hydrogen peroxide catalysed
by oxidized carbon materials. Nat. Catal. 2018, 1 (2), 156−162.
(25) Yan, X. C.; Jia, Y.; Yao, X. D. Defects on carbons for
electrocatalytic oxygen reduction. Chem. Soc. Rev. 2018, 47 (20),
7628−7658.
(26) Alt, H.; Binder, H.; Sandstede, G. Mechanism of the
electrocatalytic reduction of oxygen on metal chelates. J. Catal.
1973, 28 (1), 8−19.
(27) Zagal, J. H. Metallophthalocyanines as catalysts in electro-
chemical reactions. Coord. Chem. Rev. 1992, 119, 89−136.
́
(28) Collman, J. P.; Devaraj, N. K.; Decreau, R. A.; Yang, Y.; Yan, Y.-
L.; Ebina, W.; Eberspacher, T. A.; Chidsey, C. E. D. A cytochrome c
oxidase model catalyzes oxygen to water reduction under rate-limiting
electron flux. Science 2007, 315 (5818), 1565.
(29) Singh, K.; Razmjooei, F.; Yu, J.-S. Active sites and factors
influencing them for efficient oxygen reduction reaction in metal-N
coordinated pyrolyzed and non-pyrolyzed catalysts: a review. J. Mater.
Chem. A 2017, 5 (38), 20095−20119.
̀
(10) Lefevre, M.; Proietti, E.; Jaouen, F.; Dodelet, J.-P. Iron-based
catalysts with improved oxygen reduction activity in polymer
electrolyte fuel cells. Science 2009, 324 (5923), 71.
́
̈
(30) Auwarter, W.; Ecija, D.; Klappenberger, F.; Barth, J. V.
Porphyrins at interfaces. Nat. Chem. 2015, 7 (2), 105−120.
(31) Li, B.-Q.; Zhang, S.-Y.; Wang, B.; Xia, Z.-J.; Tang, C.; Zhang, Q.
A porphyrin covalent organic framework cathode for flexible Zn-air
batteries. Energy Environ. Sci. 2018, 11 (7), 1723−1729.
(32) Peng, P.; Shi, L.; Huo, F.; Mi, C.; Wu, X.; Zhang, S.; Xiang, Z.
A pyrolysis-free path toward superiorly catalytic nitrogen-coordinated
single atom. Sci. Adv. 2019, 5 (8), No. eaaw2322.
(11) Wu, G.; More, K. L.; Johnston, C. M.; Zelenay, P. High-
performance electrocatalysts for oxygen reduction derived from
polyaniline, iron, and cobalt. Science 2011, 332 (6028), 443−447.
(12) Shao, M. H.; Chang, Q. W.; Dodelet, J. P.; Chenitz, R. Recent
advances in electrocatalysts for oxygen reduction reaction. Chem. Rev.
2016, 116 (6), 3594−3657.
(13) Kulkarni, A.; Siahrostami, S.; Patel, A.; Norskov, J. K.
Understanding catalytic activity trends in the oxygen reduction
reaction. Chem. Rev. 2018, 118 (5), 2302−2312.
(14) Xu, H.; Cheng, D.; Cao, D.; Zeng, X. C. A universal principle
for a rational design of single-atom electrocatalysts. Nat. Catal. 2018,
1 (5), 339−348.
(15) Gao, J.; Yang, H. b.; Huang, X.; Hung, S.-F.; Cai, W.; Jia, C.;
Miao, S.; Chen, H. M.; Yang, X.; Huang, Y.; Zhang, T.; Liu, B.
Enabling Direct H2O2 Production in Acidic Media through Rational
Design of Transition Metal Single Atom Catalyst. Chem. 2020, 6 (3),
658−674.
(16) Campos-Martin, J. M.; Blanco-Brieva, G.; Fierro, J. L.
Hydrogen Peroxide Synthesis: An Outlook Beyond the Anthraqui-
none Process. Angew. Chem., Int. Ed. 2006, 45 (42), 6962−84.
(17) Siahrostami, S.; Verdaguer-Casadevall, A.; Karamad, M.;
Deiana, D.; Malacrida, P.; Wickman, B.; Escudero-Escribano, M.;
Paoli, E. A.; Frydendal, R.; Hansen, T. W.; Chorkendorff, I.; Stephens,
I. E.; Rossmeisl, J. Enabling Direct H2O2 Production through Rational
Electrocatalyst Design. Nat. Mater. 2013, 12 (12), 1137−43.
(18) Peng, Y.; Lu, B.; Chen, S. Carbon-supported single atom
catalysts for electrochemical energy conversion and storage. Adv.
Mater. 2018, 30 (48), 1801995.
(33) Li, B.-Q.; Zhang, S.-Y.; Chen, X.; Chen, C.-Y.; Xia, Z.-J.; Zhang,
Q. One-pot synthesis of framework porphyrin materials and their
applications in bifunctional oxygen electrocatalysis. Adv. Funct. Mater.
2019, 29 (29), 1901301.
(34) Strauss, V.; Roth, A.; Sekita, M.; Guldi, D. M. Efficient energy-
conversion materials for the future: understanding and tailoring
charge-transfer processes in carbon nanostructures. Chem. 2016, 1
(4), 531−556.
(35) Greeley, J.; Stephens, I. E. L.; Bondarenko, A. S.; Johansson, T.
P.; Hansen, H. A.; Jaramillo, T. F.; Rossmeisl, J.; Chorkendorff, I.;
Norskov, J. K. Alloys of platinum and early transition metals as oxygen
reduction electrocatalysts. Nat. Chem. 2009, 1 (7), 552−556.
(36) Zagal, J. H.; Koper, M. T. M. Reactivity descriptors for the
activity of molecular MN4 catalysts for the oxygen reduction reaction.
Angew. Chem., Int. Ed. 2016, 55 (47), 14510−14521.
(37) Sun, Y.; Silvioli, L.; Sahraie, N. R.; Ju, W.; Li, J.; Zitolo, A.; Li,
S.; Bagger, A.; Arnarson, L.; Wang, X.; Moeller, T.; Bernsmeier, D.;
Rossmeisl, J.; Jaouen, F.; Strasser, P. Activity-selectivity trends in the
electrochemical production of hydrogen peroxide over single-site
metal-nitrogen-carbon catalysts. J. Am. Chem. Soc. 2019, 141 (31),
12372−12381.
(19) Fei, H.; Dong, J.; Feng, Y.; Allen, C. S.; Wan, C.; Volosskiy, B.;
Li, M.; Zhao, Z.; Wang, Y.; Sun, H.; An, P.; Chen, W.; Guo, Z.; Lee,
C.; Chen, D.; Shakir, I.; Liu, M.; Hu, T.; Li, Y.; Kirkland, A. I.; Duan,
X.; Huang, Y. General synthesis and definitive structural identification
of MN4C4 single-atom catalysts with tunable electrocatalytic activities.
Nat. Catal. 2018, 1 (1), 63−72.
(38) Primbs, M.; Sun, Y.; Roy, A.; Malko, D.; Mehmood, A.;
Sougrati, M.-T.; Blanchard, P.-Y.; Granozzi, G.; Kosmala, T.; Daniel,
G.; Atanassov, P.; Sharman, J.; Durante, C.; Kucernak, A.; Jones, D. J.;
Jaouen, F.; Strasser, P. Establishing reactivity descriptors for platinum
group metal (PGM)-free Fe-N-C catalysts for PEM fuel cells. Energy
Environ. Sci. 2020, 13 (8), 2480−2500.
(20) Jiang, K.; Back, S.; Akey, A. J.; Xia, C.; Hu, Y.; Liang, W.;
Schaak, D.; Stavitski, E.; Nørskov, J. K.; Siahrostami, S.; Wang, H.
Highly selective oxygen reduction to hydrogen peroxide on transition
metal single atom coordination. Nat. Commun. 2019, 10 (1), 3997.
(21) Yang, Y.; Yang, Y.; Pei, Z.; Wu, K.-H.; Tan, C.; Wang, H.; Wei,
L.; Mahmood, A.; Yan, C.; Dong, J.; Zhao, S.; Chen, Y. Recent
progress of carbon-supported single-atom catalysts for energy
conversion and storage. Matter 2020, 3, 1442−1476.
(39) Grimaud, A.; Diaz-Morales, O.; Han, B.; Hong, W. T.; Lee, Y.-
L.; Giordano, L.; Stoerzinger, K. A.; Koper, M. T. M.; Shao-Horn, Y.
Activating lattice oxygen redox reactions in metal oxides to catalyse
oxygen evolution. Nat. Chem. 2017, 9, 457−465.
(40) Lin, C.-Y.; Zhang, L.; Zhao, Z.; Xia, Z. Design principles for
covalent organic frameworks as efficient electrocatalysts in clean
energy conversion and green oxidizer production. Adv. Mater. 2017,
29 (17), 1606635.
(41) He, Y. H.; Hwang, S.; Cullen, D. A.; Uddin, M. A.; Langhorst,
L.; Li, B. Y.; Karakalos, S.; Kropf, A. J.; Wegener, E. C.; Sokolowski, J.;
Chen, M. J.; Myers, D.; Su, D.; More, K. L.; Wang, G. F.; Litster, S.;
Wu, G. Highly active atomically dispersed CoN4 fuel cell cathode
catalysts derived from surfactant-assisted MOFs: carbon-shell confine-
ment strategy. Energy Environ. Sci. 2019, 12 (1), 250−260.
(22) Liu, X.; Dai, L. M. Carbon-based metal-free catalysts. Nat. Rev.
Mater. 2016, 1, 16082.
(23) Kim, H. W.; Ross, M. B.; Kornienko, N.; Zhang, L.; Guo, J.;
Yang, P.; McCloskey, B. D. Efficient hydrogen peroxide generation
using reduced graphene oxide-based oxygen reduction electro-
catalysts. Nat. Catal. 2018, 1 (4), 282−290.
J
https://dx.doi.org/10.1021/jacs.0c10636
J. Am. Chem. Soc. XXXX, XXX, XXX−XXX

Journal of the American Chemical Society
pubs.acs.org/JACS
Article
(42) Li, F.; Bu, Y.; Han, G.-F.; Noh, H.-J.; Kim, S.-J.; Ahmad, I.; Lu,
Y.; Zhang, P.; Jeong, H. Y.; Fu, Z.; et al. Identifying the structure of
Zn-N2 active sites and structural activation. Nat. Commun. 2019, 10
(1), 2623.
(43) Ding, S.-Y.; Wang, W. J. C. S. R. Covalent organic frameworks
(COFs): from design to applications. Chem. Soc. Rev. 2013, 42 (2),
548−568.
(44) Kohn, W.; Sham, L. J. Self-Consistent Equations Including
Exchange and Correlation Effects. Phys. Rev. 1965, 140 (4A), A1133−
A1138.
(45) Perdew, J. P.; Burke, K.; Ernzerhof, M. Generalized Gradient
Approximation Made Simple. Phys. Rev. Lett. 1996, 77 (18), 3865−
3868.
(46) Kresse, G.; Furthmuller, J. Efficient iterative schemes for ab
̈
initio total-energy calculations using a plane-wave basis set. Phys. Rev.
B: Condens. Matter Mater. Phys. 1996, 54 (16), 11169−11186.
̈
(47) Blochl, P. E. Projector augmented-wave method. Phys. Rev. B:
Condens. Matter Mater. Phys. 1994, 50 (24), 17953−17979.
́
(48) Henkelman, G.; Uberuaga, B. P.; Jonsson, H. A climbing image
nudged elastic band method for finding saddle points and minimum
energy paths. J. Chem. Phys. 2000, 113 (22), 9901−9904.
(49) Nørskov, J. K.; Rossmeisl, J.; Logadottir, A.; Lindqvist, L.;
́
Kitchin, J. R.; Bligaard, T.; Jonsson, H. Origin of the Overpotential for
Oxygen Reduction at a Fuel-Cell Cathode. J. Phys. Chem. B 2004, 108
(46), 17886−17892.
(50) Hammer, B.; Hansen, L. B.; Nørskov, J. K. Improved
adsorption energetics within density-functional theory using revised
Perdew-Burke-Ernzerhof functionals. Phys. Rev. B: Condens. Matter
Mater. Phys. 1999, 59 (11), 7413−7421.
(51) Uribe-Romo, F. J.; Hunt, J. R.; Furukawa, H.; Klock, C.;
O’Keeffe, M.; Yaghi, O. M. A crystalline imine-linked 3-D porous
covalent organic framework. J. Am. Chem. Soc. 2009, 131 (13), 4570−
4571.
(52) Lin, S.; Diercks, C. S.; Zhang, Y.-B.; Kornienko, N.; Nichols, E.
M.; Zhao, Y.; Paris, A. R.; Kim, D.; Yang, P.; Yaghi, O. M. Covalent
organic frameworks comprising cobalt porphyrins for catalytic CO2
reduction in water. Science 2015, 349 (6253), 1208−1213.
́
(53) Wan, S.; Gandara, F.; Asano, A.; Furukawa, H.; Saeki, A.; Dey,
S. K.; Liao, L.; Ambrogio, M. W.; Botros, Y. Y.; Duan, X.; et al.
Covalent organic frameworks with high charge carrier mobility. Chem.
Mater. 2011, 23 (18), 4094−4097.
(54) Gong, Y.-N.; Zhong, W.; Li, Y.; Qiu, Y.; Zheng, L.; Jiang, J.;
Jiang, H.-L. Regulating Photocatalysis by Spin-State Manipulation of
Cobalt in Covalent Organic Frameworks. J. Am. Chem. Soc. 2020, 142
(39), 16723−16731.
(55) Zitolo, A.; Ranjbar-Sahraie, N.; Mineva, T.; Li, J.; Jia, Q.;
Stamatin, S.; Harrington, G. F.; Lyth, S. M.; Krtil, P.; Mukerjee, S.;
Fonda, E.; Jaouen, F. Identification of catalytic sites in cobalt-
nitrogen-carbon materials for the oxygen reduction reaction. Nat.
Commun. 2017, 8 (1), 957.
(56) Paulus, U.; Schmidt, T.; Gasteiger, H.; Behm, R. Oxygen
reduction on a high-surface area Pt/Vulcan carbon catalyst: a thin-film
rotating ring-disk electrode study. J. Electroanal. Chem. 2001, 495 (2),
134−145.
(57) Shinagawa, T.; Garcia-Esparza, A. T.; Takanabe, K. Insight on
Tafel slopes from a microkinetic analysis of aqueous electrocatalysis
for energy conversion. Sci. Rep. 2015, 5, 13801.
(58) Viswanathan, V.; Hansen, H. A.; Rossmeisl, J.; Nørskov, J. K.
Unifying the 2e− and 4e− reduction of oxygen on metal surfaces. J.
Phys. Chem. Lett. 2012, 3 (20), 2948−2951.
(59) Amiinu, I. S.; Liu, X.; Pu, Z.; Li, W.; Li, Q.; Zhang, J.; Tang, H.;
Zhang, H.; Mu, S. From 3D ZIF nanocrystals to Co-Nx/C nanorod
array electrocatalysts for ORR, OER, and Zn-air batteries. Adv. Funct.
Mater. 2018, 28 (5), 1704638.
̈
(60) Siahrostami, S.; Bjorketun, M. E.; Strasser, P.; Greeley, J.;
Rossmeisl, J. Tandem cathode for proton exchange membrane fuel
cells. Phys. Chem. Chem. Phys. 2013, 15 (23), 9326−9334.
K
https://dx.doi.org/10.1021/jacs.0c10636
J. Am. Chem. Soc. XXXX, XXX, XXX−XXX